Steve Ballmer had an analyst review this last week and one of the top stories covered is Ballmer discussing the mistakes made focusing on Windows Vista.
I have worked on Client OS from 1988 to 2001 at Apple and Microsoft. After Windows XP in 2002 came the next version of Windows code name "Longhorn", aka Vista. There were political battles on who would head up the developer relationships for the platform and I threw in the towel before the fight to Vic Gundotra (Director of Windows Developer Relations) who now is Sr VP of Eng at Google.
Why did I turn my back on Windows Vista in 2001? After 10 years at Microsoft, I had learned my lessons, kept my mouth shut and just move on. What would have I said?
Windows Longhorn will fail because management has told the OS team to innovate. Everyone needs to be innovative. You cannot tell a whole OS team to innovate and expect it to work. The kernel team, the graphics system, printing team are all creating new ways to do things, and a simple thing like print will not work. This is what happened when Apple created the Pink OS (taligent) and faced the reality of Blue (System 7) was more realistic to ship.
Bill Gates is quoted as being surprised that Taligent was insignificant in the industry.
In 1997, when Bill Gates was asked what trend or development over the past 20 years had really caught him by surprise, his reply was:
Windows Vista was also despite 5 years of painful development shipping in Jan 2007 insignificant in the industry. While this was all going on I switched to Server OS team, then management tools, and left the company in 2006. I never touched a Windows Vista release (2002 -2007) until 2008 when I got a new laptop and it ran Vista.
The mistake made by so many executives when ordering innovation is they don't understand how the pieces work together. The team that needs to innovate works best if they know parts of the system will not change. If you try to build a new innovative building with a ground breaking foundation, state of the art super structure, with never before used electrical and mechanical systems, and new finishing material, you would need to build the building at least 3 times to get things to work.
There is a way out of this predicament. If you design the process to have many iterations of small changes with the teams working closely together it is possible. But, this team should be as small as possible and as far away from executives with top down edicts. Kind of like a SkunkWorks. When you end up with the smaller team you can make the decisions on where the innovation needs to come from and not everyone should innovate.
The mistake made is when you measure the success of a team based on how innovative they are, you have people taking risks they don't need to.
Windows Vista failed because the top down order was everyone needs to innovate to build the new version of Windows.
What end users and the media think is innovative is many times the result of many iterations of failure. Windows Vista failure. Windows 7 success. Windows 8 failure. Next OS?????